Intention-to-treat versus as-treated versus per-protocol approaches to analysis

Randomized controlled trials (RCTs) are considered the most rigorous study design for testing hypotheses and the gold standard for evaluating intervention effectiveness. However, RCTs are often conducted under the assumption of ideal conditions that may differ from real-world scenarios in which various issues, such as loss to follow-up, mistakes in participant enrollment or intervention, and low subject compliance or adherence, may occur. There are various group-defining strategies for analyzing RCT data, including the intention-to-treat (ITT), as-treated, and per-protocol (PP) approaches. The ITT principle involves analyzing all participants according to their initial group assignments, regardless of study completion and compliance or adherence to treatment protocols. This approach aims to replicate real-world clinical settings in which several anticipated or unexpected conditions may occur with regard to the study protocol. For the PP approach, only participants who meet the inclusion criteria, complete the interventions according to the study protocols, and have primary outcome data available are included. This approach aims to confirm treatment effects under optimal conditions. In general, the ITT principle is preferred for superiority and inequality trials, whereas the PP approach is preferred for equivalence and non-inferiority trials. However, both analytical approaches should be conducted and their results compared to determine whether significant differences exist. Overall, using both the ITT and PP approaches can provide a more complete picture of the treatment effects and ensure the reliability of the trial results.